Key Highlights
- $30 million investment to build a new AI factory (MC-2) in Melbourne, opening January 2026.
- Cutting-edge infrastructure powered by AMD Instinct MI355 GPUs, EPYC 9575F CPUs and ROCm software to design task-specific AI models.
- Strategic shift from general-purpose large language models (LLMs) to precision-engineered applied AI built from scratch.
In a landmark move for Australia’s artificial intelligence industry, Maincode has announced a $30 million investment to construct what it calls the country’s most advanced AI factory, located in Melbourne. It is scheduled to open in January 2026.
Known internally as “MC-2,” the facility will aim to advance Maincode’s capability to design, build and operate advanced, task-specific AI models.
Unlike traditional general-purpose LLMs, MC-2 will focus on structured, rule-based, and context-sensitive workloads, addressing specific business needs where mainstream AI models often fall short.
“MC-2 is more than an expansion, it’s a statement of belief,” said Dave Lemphers, Co-Founder and CEO of Maincode.
Beyond Hardware: Accelerating the Token Layer
At the heart of MC-2 lies cutting-edge AI infrastructure powered by AMD technology, including Instinct MI355 accelerators, EPYC 9575F CPUs, and the ROCm software stack. This combination enables high-performance training and efficient inference within a tightly integrated, scalable environment designed for both reliability and rapid deployment.
Lemphers explained that Maincode is focused not on chasing ever-bigger models, but on building applied, high-performance systems that surpass what generic LLMs can achieve. Their focus is on the “token layer”, the level where intelligence truly emerges.
This approach supports precision-engineered AI models tailored to specific domains, moving beyond generic AI systems that often lack contextual depth or operational control.
From Frustration to Factory
Many of Maincode’s partners initially experimented with off-the-shelf large language models, only to encounter challenges when they needed more control, precision, and efficiency.
The new AI factory aims to resolve those issues by creating an ecosystem where organizations can build, train, and deploy custom models faster and more effectively.
With MC-2, Maincode will bring together compute, data, research and deployment into one continuous pipeline, powering the upcoming “Matilda” models, customer-specific AI systems and advanced applied research programs.
Building the Future of Applied AI Technology
With one of the largest private-sector AI infrastructure investments in Australia, Maincode is reinforcing its role as a pioneer in applied artificial intelligence.
“What we’re building with MC-2 is capability, ” said Lemphers. “We’re using the best technology in the world and applying it with Australian inventiveness. We’re not a chatbot company; we’re an AI company that builds entirely new models from scratch, systems designed to handle the work that general-purpose models can’t, no matter how much prompt engineering you try. MC-2 isn’t about scale, it’s about precision. It’s where the next generation of useful, usable AI gets made.”
MC-2 at a Glance
- Investment: $30 million
- Facility: MC-2 – Melbourne AI Factory
- Opening Date: January 2026
- Technology Stack: AMD Instinct MI355 GPUs, AMD EPYC 9575F CPUs, ROCm software
- Focus: Advanced, task-specific AI models built from the ground up
- Flagship Model: Matilda – Australia’s first fully built and trained large language model
This $30 million bet on AI infrastructure signals a bold new era for Australian technology, positioning Maincode at the forefront of the applied AI revolution.